On the Effectiveness of Dataset Embeddings in Mono-lingual, Multi-lingual and Zero-shot Conditions

Keyphrases

• Performance Increase 100%
• Heterogeneous Data Sources 33%
• Dependency Parsing 33%
• Lemmatization 33%
• Morphosyntax 33%
• Single Model 33%
• Predicted Data 33%
• Multilingual Context 33%
• Morphological Tagging 33%
• Distribution Performance 33%
• Complementary Strand 33%
• Grouping Strategy 33%
• Different Datasets 33%
• Unseen 33%

Social Sciences

• Parsing 100%
• Lemmatization 100%

Computer Science

• Data Source 100%
• Heterogeneous Data 25%
• dependency parse 25%